In Search of an Integrative Method to Study Unconscious Processing: An Application of Bayesian and General Recognition Theory Models to the Processing of Hierarchical Patterns in the Absence of Awareness

Antonio Prieto; Pedro R. Montoro; Mikel Jimenez; José Antonio Hinojosa

doi:10.5334/joc.411

Figures & Tables

Stimuli used in the four experiments. **(a)** Prime stimuli for global priming experiments. **(b)** Prime stimuli for local priming experiments. **(c)** Probe stimuli for all the experiments. **(d)** Examples of mask stimuli.

Sequence of events in each block of the experiments.

Table 1

Priming condition and variable display durations (ISI and SOA) in ms across experiments.

EXPERIMENT	PRIMING CONDITION	VARIABLE DISPLAY DURATIONS		PRIME-TARGET SOA
EXPERIMENT	PRIMING CONDITION	ISI	BACKWARD MASK	PRIME-TARGET SOA
1	Global	0	67	107
2	Global	13	53	107
3	Local	0	67	107
4	Local	13	53	107

Priming effects (Incongruent – Congruent) in both the single-task priming block, the dual-task priming block (unfiltered), and the dual-task priming block (filtered by PAS1). Error bars represent standard error (SE) of the mean.

Table 2

Mean (M) and standard error (SE) for RTs (ms) in congruent and incongruent conditions in the single-task priming block, multiple-task priming block (unfiltered), and the multiple-task priming block (filtered by PAS1) across all four experiments.

TASK	RT
	CONGRUENT		INCONGRUENT
	M	SE	M	SE
Single
Global 40 ms	518	9.9	527	10.4
Global 53 ms	548	14.3	554	12.7
Local 40 ms	520	13.3	519	13.6
Local 53 ms	569	20.9	565	20
Multiple
Global 40 ms	780	42.4	786	43.5
Global 53 ms	859	46.3	870	46.4
Local 40 ms	858	46.2	888	53
Local 53 ms	951	53	983	60.3
Multiple PAS-1
Global 40 ms	765	40	770	40.7
Global 53 ms	852	46.1	877	49.9
Local 40 ms	814	41.15	828	42.73
Local 53 ms	957	64.1	889	53.15

Depiction of the Bayesian correction to Greenwald regression in Experiment 1 (Global SOA 40) in the single-task **(A)** and multiple-task **(B)** blocks. Each circle represents the participant estimated true score. The x-axis represents the centered awareness score, the y-axis represents the estimated effect (incongruent – congruent conditions). The intercept is the expected performance for completely unaware participants.

Table 3

Mean (M) and standard error (SE) for PAS reports (%) in the multiple-task priming block and the prime visibility block across all four experiments.

PAS	EXPERIMENT
PAS	GLOBAL 40 MS		GLOBAL 53 MS		LOCAL 40 MS		LOCAL 53 MS
BLOCK	M	SE	M	SE	M	SE	M	SE
Multiple-task
1	70.5	6	65	6.9	52.7	6.3	51.4	8.4
2	19.8	4	22	4.6	21.7	3	19	4
3	8.4	2.6	7.6	2.8	14.8	3.3	12.6	3.4
4	1.4	1.1	5.4	3.7	10.8	3.7	17	5.6
Prime Visibility
1	64.1	6.5	52.3	7.6	17.2	5.1	18.8	5.6
2	18.7	3.4	25.2	4.1	20.9	3.6	22.2	5.4
3	11.2	2.7	16.3	4.4	24.9	4.2	22.1	4.7
4	6	2.9	6.2	3.3	37	7.5	36.9	7.5

Depiction of the Bayesian correction to Greenwald regression in Experiment 2 (Global SOA 53) in the single-task **(A)** and multiple-task **(B)** blocks. Each circle represents the participant estimated true score. The x-axis represents the centered awareness score, the y-axis represents the estimated effect (incongruent – congruent conditions). The intercept is the expected performance for completely unaware participants.

Depiction of the Bayesian correction to Greenwald regression in Experiment 3 (Local SOA 40) in the single-task **(A)** and multiple-task **(B)** blocks. Each circle represents the participant estimated true score. The x-axis represents the centered awareness score, the y-axis represents the estimated effect (incongruent – congruent conditions). The intercept is the expected performance for completely unaware participants.

Depiction of the Bayesian correction to Greenwald regression in Experiment 4 (Local SOA 53) in the single-task **(A)** and multiple-task **(B)** blocks. Each circle represents the participant estimated true score. The x-axis represents the centered awareness score, the y-axis represents the estimated effect (incongruent – congruent conditions). The intercept is the expected performance for completely unaware participants.

Sensitivity vs. awareness (SvA) curves obtained during the Experiment 1 (Global SAO 40 during the multiple-task **(A)** and visibility blocks **(B)**. The solid red line represents the SvA curve obtained from the best adjusted model, and the lighter red bands represents 95% confidence intervals. The dotted blue line separates regions of relative high likelihood of awareness to the left, and relative low regions o awareness to the right. The dotted green lines are the estimated bounds for each participant. The horizontal black dotted line represents zero sensitivity (d’ = 0) in the prime shape discrimination task.

Sensitivity vs. awareness (SvA) curves obtained during the Experiment 2 (Global SOA 53) during the multiple-task **(A)** and visibility blocks **(B)**. The solid red line represents the SvA curve obtained from the best adjusted model, and the lighter red bands represents 95% confidence intervals. The dotted blue line separates regions of relative high likelihood of awareness to the left, and relative low regions o awareness to the right. The dotted green lines are the estimated bounds for each participant. The horizontal black dotted line represents zero sensitivity (d’ = 0) in the prime shape discrimination task.

Sensitivity vs. awareness (SvA) curves obtained during the Experiment 3 (Local SOA 40) during the multiple-task **(A)** and visibility blocks **(B)**. The solid red line represents the SvA curve obtained from the best adjusted model, and the lighter red bands represents 95% confidence intervals. The dotted blue line separates regions of relative high likelihood of awareness to the left, and relative low regions o awareness to the right. The dotted green lines are the estimated bounds for each participant. The horizontal black dotted line represents zero sensitivity (d’ = 0) in the prime shape discrimination task.

Sensitivity vs. awareness (SvA) curves obtained during the Experiment 4 (Local SOA 53) during the multiple-task **(A)** and visibility blocks **(B)**. The solid red line represents the SvA curve obtained from the best adjusted model, and the lighter red bands represents 95% confidence intervals. The dotted blue line separates regions of relative high likelihood of awareness to the left, and relative low regions o awareness to the right. The dotted green lines are the estimated bounds for each participant. The horizontal black dotted line represents zero sensitivity (d’ = 0) in the prime shape discrimination task.

Sensitivity measures (d’) obtained during the multiple-task and visibility blocks (d’_obj and d’_subj).

Table 4

d’ values for objective and subjective sensitivity measures of prime shape discrimination. Bayes factor (BF₁₀) and Pearson’s correlation coefficients (r) for the relevant comparisons between sensitivity measures across all four experiments. * BF₁₀ > 10, ** BF₁₀ > 30, *** BF₁₀ > 100.

EXPERIMENT	d’ VALUES	CORRELATION	BF₁₀	PEARSON’S r
Global SOA40 ms	d’_{obj(multiple-task)} = 0.073	d’_{obj(multiple-task)} – d’_{subj(multiple-task)}	5.259	0.463
	d’_{obj(visibility)} = 0.319	d’_{obj(visibility)} – d’_{subj(visibility)}	289.221	0.694***
	d’_{subj(multiple-task)} = 0.151	d’_{obj(multiple-task)} – d’_{obj(visibility)}	1865.976	0.757***
	d’_{subj(visibility)} = 0.392	d’_{subj(multiple-task)} – d’_{subj(visibility)}	824.921	0.731***
Global SOA53 ms	d’_{obj(multiple-task)} = 0.230	d’_{obj(multiple-task)} – d’_{subj(multiple-task)}	15.906	0.625*
	d’_{obj(visibility)} = 0.515	d’_{obj(visibility)} – d’_{subj(visibility)}	1916.471	0.814***
	d’_{subj(multiple-task)} = 0.226	d’_{obj(multiple-task)} _– d’_{obj(visibility)}	163.048	0.737***
	d’_{subj(visibility)} = 0.596	d’_{subj(multiple-task)} – d’_{subj(visibility)}	25258.694	0.869***
Local SOA40 ms	d’_{obj(multiple-task)} = 0.575	d’_{obj(multiple-task)} – d’_{subj(multiple-task)}	56.915	0.622**
	d’_{obj(visibility)} = 2.551	d’_{obj(visibility)} – d’_{subj(visibility)}	8642.373	0.781***
	d’_{subj(multiple-task)} = 1.162	d’_{obj(multiple-task)} – d’_{obj(visibility)}	203.543	0.674***
	d’_{subj(visibility)} = 2.609	d’_{subj(multiple-task)} – d’_{subj(visibility)}	227.081	0.678***
Local SOA53 ms	d’_{obj(multiple-task)} = 0.748	d’_{obj(multiple-task)} – d’_{subj(multiple-task)}	293.053	0.720***
	d’_{obj(visibility)} = 2.105	d’_{obj(visibility)} – d’_{subj(visibility)}	47.654	0.646**
	d’_{subj(multiple-task)} = 1.395	d’_{obj(multiple-task)} – d’_{obj(visibility)}	2288.726	0.782***
	d’_{subj(visibility)} = 2.706	d’_{subj(multiple-task)} – d’_{subj(visibility)}	9865.952	0.816***

Table 5

Summary of the results according to the different types of analysis performed.

ANALYSIS METHOD	MAIN RESULTS
Classical dissociation paradigm: (Objective awareness measures)	Evidence favoring priming effects for the global shape in Experiments 1 (strong) and 2 (anecdotal) during the single-task block. Evidence against priming effects for the local elements in Experiments 3 and 4 (moderate) during the single-task block. No evidence for priming effects for the global shape in Experiments 1 and 2 during the multiple task block Evidence favoring priming effects for the local elements in Experiments 3 and 4 (moderate) during the multiple-task block. Increased RT and variability during the multiple-task block compared to the single-task block All d’ > 0 except in the multiple-task block of Experiment 1
Perceptual awareness scale (PAS-1): (Subjective awareness measures)	Evidence against the existence of unconscious priming effects for the global shape in the multiple-task block of Experiment 1 (moderate) and 2 (anecdotal) Evidence against the existence of unconscious priming effects for the local elements in the multiple-task block of Experiment 3 (anecdotal) and 4 (anecdotal). No evidence of priming effects when only PAS-1 (non-conscious) trials were analyzed
Bayesian regression (Objective awareness measures)	Evidence against the existence of unconscious priming effects for the global shape in the single-task block of Experiment 1 (moderate) and 2 (moderate). Evidence against the existence of unconscious priming effects for the local elements in the single-task block of Experiment 3 (strong) and 4 (strong). Evidence against the existence of unconscious priming effects for the global shape in the multiple-task block of Experiment 1 (moderate) and 2 (moderate). Evidence against the existence of unconscious priming effects for the local elements in the multiple-task block of Experiment 3 (strong) and 4 (moderate).
GRT-based SvA curves (Subjective awareness measures)	Greater than 0 sensitivity (d’) when RLNA was high (>1) for the global shapes in Experiment 1 (≈ 0.25) and 2 (≈ 0.6) during the multiple-task block Greater than 0 sensitivity (d’) when RLNA was high (>1) for local elements in Experiment 3 (≈ 0.75) and 4 (≈ 1.7) during the multiple-task block Greater than 0 sensitivity (d’) when RLNA was high (>1) for the global shapes in Experiment 1 (≈ 1.25) and 2 (≈ 1.5) during the visibility block Greater than 0 sensitivity (d’) when RLNA was high (>1) for local elements in Experiment 3 (≈ 2.5) and 4 (≈ 2.7) during the visibility block
Objective VS subjective awareness measures comparison (d’_obj vs d’_subj)	Greater d’ values (both objective and subjective during the visibility block (full attention to the primes), compared to the multiple-task block (divided attention between primes and probes) in all experiments. Overall greater d’ values when assessed by means of subjective awareness measures, compared to objective awareness measures. Overall strong correlations between objective (d’_obj) and subjective (d’_subj) measures of awareness within the same block (multiple-task and visibility blocks respectively) Overall strong correlations within objective (d’_obj) measures of awareness collected in different blocks (multiple-task and visibility blocks respectively) Overall strong correlations within subjective (d’_subj) measures of awareness collected in different blocks (multiple-task and visibility blocks respectively)

In Search of an Integrative Method to Study Unconscious Processing: An Application of Bayesian and General Recognition Theory Models to the Processing of Hierarchical Patterns in the Absence of Awareness

Figures & Tables

Figure 1

Figure 2

Table 1

Figure 3

Table 2

Figure 4

Table 3

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Table 4

Table 5

Paradigm

My account